[Community] [Gta04-owner] OpenPhoenux Letux/GTA04 Kernel- and Userspace-Hacking weekend
H. Nikolaus Schaller
hns at goldelico.com
Mon Apr 25 11:54:19 CEST 2016
> Am 21.04.2016 um 18:40 schrieb Neil Jerram <neil at ossau.homelinux.net>:
> On 21/04/16 15:33, H. Nikolaus Schaller wrote:
>> Hi all,
>> finally, we have now defined the location and weekend dates:
>> 4 & 5 June 2016 in Munich
>> Details about the location are updated here:
>> There will also be a pre-meeting "Stammtisch" open to everybody on 3rd June. So please
>> make your travel arrangements. If you need support, please contact me.
>> For best preparation, I think we should have these things ready when the workshop starts:
>> * everyone has a notebook with required compilers, source codes, tools etc.
>> * and (at least one) device to test on with cable/SD card flashing tools
>> The idea is that we are maximally productive and not wasting (too much) time to setup hardware.
>> WLAN is available at the site as well as a beamer for short presentations or code reviews in a group.
>> Especially for the QtMoko task it would be good if someone could set up a build environment
>> that works. In my experiments it did take several days (independently of trying virtual machines
>> or OMAP5 hardware) but finally failed (too big or missing dependencies).
>> It would also be nice if we could collect material (slide show?) in advance about the architecture
>> of QtMoko and how the build system should work, so that we can share this knowledge in the
>> workshop warmup.
>> And there is a question I want to rise in advance: and wihc should have been answered when
>> we begin to work.
>> What is easier:
>> a) porting existing complete QtMoko/QPE source and build system to build and run on Debian Jessie
>> b) or rewriting/porting just the most important QtMoko applications to Qt5 (which is
>> easily available in Debian Jessie and works within minutes on a GTA04 - I have tested)?
> I personally favour (b), because I want my GTA04 system to be as close to a normal Linux as possible, including being able to mix apps of all kinds: X, Gtk, Qt and so on.
Yes, that was the line of thought when I came to the idea to propose b). Just make Qt a set of debian bundles and a meta package to install everything (or groups).
> Rewriting the interesting parts of QtMoko so as to work with a normal desktop environment would be a great help with that.
Exactly. It more seamlessly integrates other components and since Qt4/5 have different backends (X11) it would be easier to port to newer and more devices. E.g. Pyra.
> However, I'm afraid I have no idea if it would be easy, or easier than (a).
That is really difficult to judge... I have no idea how difficult it is to get (a) working again. There may be a lot of hidden obstacles making QPE build on newer OS libraries and environments.
On the other hand, rewriting the ca. 50 Apps (http://qtmoko.sourceforge.net/apps/) from scratch and more importantly test them is also a lot of work. So we might be fatster to get to something useable with path (a) and have only to iron out bugs that appear in daily use.
We could also try the tools to semi-automatically update Qt3 projects to Qt4/5. The main differences appear to be changes in include paths and renamed/(re)moved Qt classes. But we can keep the general architecture and code structure which is of great help.
> Also BTW I believe there is lots of QtMoko NIH-ware that is now obsoleted by better components in general Linux. For example, connman or network-manager instead of QtMoko's own networking management, and offlineimap or mbsync instead of QtMoko's own IMAP mailstore.
Well, in my experience such existing tools are rarely optimized for touch and finger operation on small vertical screens. So we either have to tweak them or write something new. Or port something old.
That is IMHO a strong point of QtMoko: that it is more or less self-contained for the basic operation without the need to harmonize GUI operation between tools written with Desktop in mind.
But I agree that it should interwork with other apps from the standard Debian repositories.
> The interesting parts of QtMoko for me are its UI, such as its launcher, phone and SMS UIs, and its on-screen keyboard.
Rewriting launcher, phone and SMS UI should be not impossible tasks.
But I have no idea if the on-screen keyboard can be rewritten in a way that it works with all other GUI applications (not necessarily Qt based!). You would get a problem if you have network-manager and can't type IP addresses... So this might be a big challenge (who doesn't love challenges?).
Any thoughts about this from the community?
Finally I have a (late) easter egg for us: http://www.qtmoko.net
More information about the Community